Handle ghost Aquatroll tables at OWC; add OWC to testing data

[bpbond]

I experimented with changing the default column type to character, but that caused problems elsewhere. So instead in read_datalogger_file() we now ensure that any column named "Aquatroll_ID..." has been read as character:

    # Some Aquatroll tables (specifically at OWC; see e.g. #21) have blank
    # tables, and their "Aquatroll_IDx" columns get parsed as numeric, which
    # causes errors at the bind_rows() step below. Pre-emptively fix this
    fixcol <- grep("^Aquatroll_ID", colnames(x))
    if(length(fixcol) > 0) x[fixcol] <- as.character(x[fixcol])

A little hacky ☹️ but it works.

Closes #21

[bpbond]

See https://rpubs.com/bpbond/1330934 for an example of plotting OWC 🎉

[wilsonsj100]

YAY! Thank you @bpbond I appreciate you decoding all these issues :) I am going to merge then try it on my own computer